
import csv
import snscrape.modules.twitter as sntwitter

max_len = 30_000


def save_to_csv(name, tweets):
    fieldnames = ['ID', 'Content', 'User', 'Date', 'Lang', 'Place', 'Hashtags']
    with open(name, 'w', newline='', encoding='utf-8') as file:
        writer = csv.DictWriter(file, fieldnames)
        writer.writeheader()
        for tweet in tweets:
            writer.writerow({'ID': tweet.id,
                             'Content': str(tweet.rawContent).replace('\n', ''),
                             'User': tweet.user.username,
                             'Date': tweet.date,
                             'Lang': tweet.lang,
                             'Place': tweet.place,
                             'Hashtags': tweet.hashtags})

    print(f"Pobrane tweety zostały zapisane do pliku: {name}")


def q1():
    query = '(#cyberpunk2077) until:2018-06-12 since:2018-06-10'
    tweets = []
    i = 0

    for tweet in sntwitter.TwitterSearchScraper(query).get_items():
        tweets.append(tweet)
        i += 1
        print(i, tweet.date, query)

        if len(tweets) >= max_len:
            break

    csv_file = 'data_MIXED/cyberpunk_first_trailer.csv'
    save_to_csv(csv_file, tweets)


from googletrans import Translator


def translate_to_english(text, lang):
    if lang == 'en':
        return text
    elif lang == 'und' or lang == 'qme' or lang == 'qht':
        return text
    elif lang == 'zh':
        translator = Translator(service_urls=['translate.googleapis.com'])
        translation = translator.translate(text, dest='en', src='zh-tw')
        return translation.text
    elif lang == 'in':
        translator = Translator(service_urls=['translate.googleapis.com'])
        translation = translator.translate(text, dest='en', src='es')
        return translation.text
    else:
        translator = Translator(service_urls=['translate.googleapis.com'])
        translation = translator.translate(text, dest='en', src=lang)
        return translation.text
    

def q1():
    translate_to_english.call_counter = 0
    df = pd.read_csv('data_MIXED/cyberpunk_first_move.csv')
    df['Content'] = df.apply(lambda row: translate_to_english(row['Content'], row['Lang']), axis=1)

    df.to_csv('ENG_first_move.csv', index=False)


import nltk
import re
import csv
from nltk.tokenize import word_tokenize
from nltk.corpus import stopwords
from nltk.stem import WordNetLemmatizer


nltk.download('punkt')
nltk.download('stopwords')
nltk.download('wordnet')

def preprocessing(nazwa,phrases_to_remove,words_to_remove): # funkcja zajmująca się preprocessingiem
    lemmatizer = WordNetLemmatizer()
    stop_words = set(stopwords.words('english'))

    tweets = []
    with open(nazwa, 'r',encoding='utf-8') as file:
        reader = csv.DictReader(file)
        for row in reader:
            tweet = row['Content']
            tweet = re.sub(r'http\S+|https\S+', '', tweet) # usuwanie linków z tweetów
            for phrase in phrases_to_remove:
                tweet = tweet.replace(phrase, '') # usunięcie powtarzajacych się fraz
            tokens = word_tokenize(tweet) # tokenizacja
            tokens = [token.lower() for token in tokens if token.isalpha()]  # Usunięcie znaków niebędących literami
            tokens = [token for token in tokens if token not in stop_words]  # Usunięcie stopwords
            tokens = [lemmatizer.lemmatize(token) for token in tokens]  # Lematyzacja
            tokens = [token for token in tokens if token not in words_to_remove]# Usunięcie pojedynczych słów

            processed_tweet = ' '.join(tokens)
            tweets.append(processed_tweet)
    return tweets

phrases_to_remove = ['cyberpunk', 'youtube', 'video', 'game', 'liked youtube','youtube cyberpunk','cyberpunk official','official trailer','official trailer cyberpunk','liked youtube','cyberpunk official','liked youtube cyberpunk',"halo infinite","dying","light","devil","cry", "devil may cry", "dying light","shadow die","fallout","metro exodus","tomb raider","kingdom heart","forza horizon","microsoft","division","xbox","forza horizon"]
words_to_remove = ['may', 'cry','cyberpunk','youtube','shadow','devil','die','dying','trailer','gear','war']

tweets_first_trailer = preprocessing('data_ENGLISH/ENG_first_trailer.csv',phrases_to_remove,words_to_remove)
tweets_first_gameplay = preprocessing('data_ENGLISH/ENG_first_gameplay.csv',phrases_to_remove,words_to_remove)
tweets_first_move = preprocessing('data_ENGLISH/ENG_first_move.csv',phrases_to_remove,words_to_remove)
tweets_second_move = preprocessing('data_ENGLISH/ENG_second_move.csv',phrases_to_remove,words_to_remove)
tweets_third_move = preprocessing('data_ENGLISH/ENG_third_move.csv',phrases_to_remove,words_to_remove)
tweets_before_release = preprocessing('data_ENGLISH/ENG_before_release.csv',phrases_to_remove,words_to_remove)
tweets_after_release = preprocessing('data_ENGLISH/ENG_after_release.csv',phrases_to_remove,words_to_remove)

[nltk_data] Downloading package punkt to
[nltk_data]     C:\Users\26kar\AppData\Roaming\nltk_data...
[nltk_data]   Package punkt is already up-to-date!
[nltk_data] Downloading package stopwords to
[nltk_data]     C:\Users\26kar\AppData\Roaming\nltk_data...
[nltk_data]   Package stopwords is already up-to-date!
[nltk_data] Downloading package wordnet to
[nltk_data]     C:\Users\26kar\AppData\Roaming\nltk_data...
[nltk_data]   Package wordnet is already up-to-date!


import csv
import re
import nltk
from wordcloud import WordCloud
from matplotlib import pyplot as plt
from nltk.tokenize import word_tokenize
from nltk.corpus import stopwords
from nltk.stem import WordNetLemmatizer
from nltk.sentiment import SentimentIntensityAnalyzer
nltk.download('vader_lexicon')

def wordClouds(tweets, name):
    wordcloud_text = ' '.join(tweets)
    wordcloud = WordCloud(width=800, height=400).generate(wordcloud_text)

    plt.figure(figsize=(10, 5))
    plt.imshow(wordcloud, interpolation='bilinear')
    plt.axis('off')
    plt.title(name)
    plt.show()


all_tweets = tweets_first_trailer + tweets_first_gameplay + tweets_first_move + tweets_second_move+ tweets_third_move+ tweets_before_release + tweets_after_release

wordClouds(all_tweets,"Ogólne słowa dotyczące Cyberpunk2077")

[nltk_data] Downloading package vader_lexicon to
[nltk_data]     C:\Users\26kar\AppData\Roaming\nltk_data...
[nltk_data]   Package vader_lexicon is already up-to-date!


from text2emotion import get_emotion
from wordcloud import WordCloud
import matplotlib.pyplot as plt


def wordClouds(tweets, name):
    wordcloud_text = ' '.join(tweets)

    wordcloud = WordCloud(width=800, height=400).generate(wordcloud_text)

    plt.figure(figsize=(10, 5))
    plt.imshow(wordcloud, interpolation='bilinear')
    plt.axis('off')
    plt.title(name)
    plt.show()

def process_tweets(all_tweets):
    happy_tweets = []
    sadness_tweets = []
    anger_tweets = []
    fear_tweets = []
    surprise_tweets = []

    for tweet in all_tweets:
        emotions = get_emotion(tweet)
        dominant_emotion = max(emotions, key=emotions.get)

        if dominant_emotion == 'Happy':
            happy_tweets.append(tweet)
        elif dominant_emotion == 'Sad':
            sadness_tweets.append(tweet)
        elif dominant_emotion == 'Angry':
            anger_tweets.append(tweet)
        elif dominant_emotion == 'Fear':
            fear_tweets.append(tweet)
        elif dominant_emotion == 'Surprise':
            surprise_tweets.append(tweet)

    wordClouds(happy_tweets,"Happy tweets")
    wordClouds(sadness_tweets,"Sad tweets")
    wordClouds(anger_tweets,"Angry tweets")
    wordClouds(fear_tweets,"Fear tweets")
    wordClouds(surprise_tweets,"Surprise tweets")


process_tweets(all_tweets)


def analiza(tweets,topic):
    sia = SentimentIntensityAnalyzer()
    positive_tweets=[]
    negative_tweets=[]
    for tweet in tweets:
        sentiment = sia.polarity_scores(tweet)
        if sentiment['compound'] >= 0.05:
            positive_tweets.append(tweet)
        elif sentiment['compound'] <= -0.05:
            negative_tweets.append(tweet)

    wordClouds(positive_tweets,'Pozytywne określenia odnośnie ' + topic)
    wordClouds(negative_tweets,'Negatywne określenia odnośnie '+ topic)


analiza(tweets_first_trailer,'pierwszego zwiastuna')


analiza(tweets_first_gameplay,'pierwszego zwiastuna rozgrywki')


analiza(tweets_first_move,'pierwszego przesunięcia daty premiery')


analiza(tweets_second_move,'drugiego przesunięcia daty premiery')


analiza(tweets_third_move,'trzeciego przesunięcia daty premiery')


analiza(tweets_before_release,'oczekiwań przed premierą')


analiza(tweets_after_release,'wrażeń po premierze')


import matplotlib.dates as mdates
from datetime import datetime
import matplotlib.pyplot as plt
import pandas as pd
import text2emotion as te


def calculate_average_emotions(files):
    emotions = ['Happy', 'Angry', 'Surprise', 'Sad', 'Fear']
    dates = []
    all_emotions = []

    for file in files:
    
        averages = {emotion: 0 for emotion in emotions}
        data = pd.read_csv(file)
        file_length = len(data)
        dates.append(data['Date'].values[0])

        for content in data['Content']:
            result = te.get_emotion(content)
            for emotion in emotions:
                averages[emotion] += result[emotion]

        for emotion in emotions:
            averages[emotion] = averages[emotion] / file_length

        all_emotions.append(averages)
    plot_emotion_changes(dates,all_emotions)


def plot_emotion_changes(dates, emotions):
    plt.figure(figsize=(20, 6))
    x = [datetime.strptime(date, "%Y-%m-%d %H:%M:%S%z") for date in dates]
    y_happy = [emotion['Happy'] for emotion in emotions]
    y_angry = [emotion['Angry'] for emotion in emotions]
    y_surprise = [emotion['Surprise'] for emotion in emotions]
    y_sad = [emotion['Sad'] for emotion in emotions]
    y_fear = [emotion['Fear'] for emotion in emotions]

    fig, ax = plt.subplots()
    ax.yaxis.tick_left()
    ax.xaxis.set_major_locator(mdates.MonthLocator())
    ax.xaxis.set_major_formatter(mdates.DateFormatter('%d-%m-%Y'))

    ax.plot(x, y_happy, label='Happy', color='g', marker='o')
    ax.plot(x, y_angry, label='Angry', color='r', marker='o')
    ax.plot(x, y_surprise, label='Surprise', color='c', marker='o')
    ax.plot(x, y_sad, label='Sad', color='b', marker='o')
    ax.plot(x, y_fear, label='Fear', color='m', marker='o')

    ax.legend(loc='center left', bbox_to_anchor=(1, 0.5))
    ax.set_xlabel('Data')
    ax.set_ylabel('Poziom Emocji')
    plt.xticks(rotation=45)

    ax.grid()
    plt.title('Analiza Czasowa Emocji')
    plt.savefig('Analiza Czasowa Emocji.png')
    plt.show()


file_paths = ['data_ENGLISH/ENG_first_trailer.csv', 'data_ENGLISH/ENG_first_gameplay.csv', 'data_ENGLISH/ENG_first_move.csv',
              'data_ENGLISH/ENG_second_move.csv', 'data_ENGLISH/ENG_third_move.csv', 'data_ENGLISH/ENG_before_release.csv', 'data_ENGLISH/ENG_after_release.csv']
calculate_average_emotions(file_paths)

[nltk_data] Downloading package stopwords to
[nltk_data]     C:\Users\26kar\AppData\Roaming\nltk_data...
[nltk_data]   Package stopwords is already up-to-date!
[nltk_data] Downloading package punkt to
[nltk_data]     C:\Users\26kar\AppData\Roaming\nltk_data...
[nltk_data]   Package punkt is already up-to-date!
[nltk_data] Downloading package wordnet to
[nltk_data]     C:\Users\26kar\AppData\Roaming\nltk_data...
[nltk_data]   Package wordnet is already up-to-date!

<Figure size 2000x600 with 0 Axes>


def calculate_average_emotions(file):
    emotions = ['Happy', 'Angry', 'Surprise', 'Sad', 'Fear']
    tweets = 0
    tweets_per_day = 300
    data = pd.read_csv(file)
    all_emotions = []

    data['Date'] = pd.to_datetime(data['Date'])
    unikalne_daty = data['Date'].dt.date.unique()

    for date in unikalne_daty:
        subset = data[data['Date'].dt.date == date]
        content = subset['Content'].tolist()
        if tweets == 35705:
            break

        for i in range(0, len(content), tweets_per_day):
            batch = content[i:i + tweets_per_day]
            averages = {emotion: 0 for emotion in emotions}
            for content in batch:
                tweets += 1
                result = te.get_emotion(content)

                for emotion in emotions:
                    averages[emotion] += result[emotion]

            for emotion in emotions:
                averages[emotion] = averages[emotion] / tweets_per_day

            all_emotions.append(averages)

    data_string = []
    for data in unikalne_daty:
        data_string.append(data.strftime('%Y-%m-%d'))

    unikalne_daty = data_string
    plot_emotion_changes(unikalne_daty,all_emotions)



def plot_emotion_changes(dates, emotions):

    x = [datetime.strptime(date, "%Y-%m-%d") for date in dates]
    y_happy = [emotion['Happy'] for emotion in emotions]
    y_angry = [emotion['Angry'] for emotion in emotions]
    y_surprise = [emotion['Surprise'] for emotion in emotions]
    y_sad = [emotion['Sad'] for emotion in emotions]
    y_fear = [emotion['Fear'] for emotion in emotions]

    fig, ax = plt.subplots(figsize=(20, 6))
    ax.yaxis.tick_left()
    ax.xaxis.set_major_locator(mdates.MonthLocator())
    ax.xaxis.set_major_formatter(mdates.DateFormatter('%d-%m-%Y'))

    ax.plot(x, y_happy, label='Happy', color='g')
    ax.plot(x, y_angry, label='Angry', color='r')
    ax.plot(x, y_surprise, label='Surprise', color='c')
    ax.plot(x, y_sad, label='Sad', color='b')
    ax.plot(x, y_fear, label='Fear', color='m')

    ax.legend(loc='center left', bbox_to_anchor=(1, 0.5))
    ax.set_xlabel('Data')
    ax.set_ylabel('Poziom Emocji')
    plt.xticks(rotation=45)

    ax.grid()
    plt.title('Analiza Czasowa Emocji od listopada 2020 do lutego 2021')
    plt.savefig('Analiza Czasowa Emocji - daily.png')
    plt.show()


file= 'data_ENGLISH/ENG_daily_tweets.csv'
calculate_average_emotions(file)


import pandas as pd
import matplotlib.pyplot as plt

df = pd.read_csv('data_MIXED/CDP-kursy.csv')

df['Data'] = pd.to_datetime(df['Data'])

plt.figure(figsize=(20, 6))
plt.plot(df['Data'], df['Zamknięcie'])
plt.xlabel('Data')
plt.ylabel('Zamknięcie')
plt.title('Wykres kursu spółki')
plt.grid(True)
plt.xticks(rotation=45)
plt.show()


import csv
import pandas as pd
from nltk.sentiment import SentimentIntensityAnalyzer
import matplotlib.pyplot as plt

sia = SentimentIntensityAnalyzer()

positive_tweets = []
negative_tweets = []

positive_lengths = []
negative_lengths = []
dates = []

with open('data_ENGLISH/ENG_daily_tweets.csv', 'r', encoding='utf-8') as file:
    reader = csv.DictReader(file)

    for i, row in enumerate(reader):
        tweet = row['Content']
        date = row['Date']
        date = date[:10]

        sentiment = sia.polarity_scores(tweet)

        if sentiment['compound'] >= 0.05:
            positive_tweets.append(tweet)
        elif sentiment['compound'] <= -0.05:
            negative_tweets.append(tweet)

        if (i + 1) % 300 == 0:
            positive_lengths.append(len(positive_tweets))
            negative_lengths.append(len(negative_tweets))
            negative_tweets = []
            positive_tweets = []
            dates.append(date)

df = pd.DataFrame({'Positive': positive_lengths, 'Negative': negative_lengths}, index=dates)

df.plot(kind='line', figsize=(20, 6))
plt.grid()
plt.title('Liczba pozytywnych i negatywnych tweetów od listopada 2020 do lutego 2021')
plt.xlabel('Data')
plt.ylabel('Liczba tweetów')
plt.legend()
plt.show()

Analiza opinii z Twittera na temat gry Cyberpunk2077¶

1. Wstęp¶

Pobieranie tweetów¶

2. Preprocessing¶

Tłumaczenie¶

Preprocessing dla Bag of Words¶

3. Analiza opinii¶

Analiza całościowa¶

Ogólne słowa związane z #cyberpunk2077¶

Słowa padające w kontekście każdej z 5 emocji związanych z Cyberpunkiem2077¶

Słowa padające w kazdym z okresów¶

Analiza czasowa według wybranych okresów¶

Analiza czasowa w czasie premiery¶

4. Podsumowanie¶

5. Bibliografia¶